Arrival Angle Estimation Device (oyobi) Arrival Angle Estimation Method

Abstract

The present invention provides an angle of arrival estimation device and method that can estimate the angle of arrival without generating or suppressing bias errors, even when there are unknown gain fluctuations. [Solution] The arrival angle estimation device of this embodiment comprises: an antenna in which at least one of the beam gain and beam shape changes with the beam angle; a power detection unit that detects the power of radio waves from a radio wave source received by the antenna based on the beam; a beam control unit that controls the beam angle; and an arrival angle estimation unit that estimates the arrival angle of the radio waves based on the power detected by the power detection unit at the beam angles at the expected arrival angle of the radio waves, swinging the beam in two or more directions from the center with different amplitudes so that the beam gain at the center is equal.

Description

Technical Field

[0001] Embodiments of the present invention relate to an arrival angle estimation device and an arrival angle estimation method, and more particularly, to an arrival angle estimation device and an arrival angle estimation method for estimating the direction of a radio wave oscillation source by, for example, swinging an antenna beam.

Background Art

[0002] In satellite communication, if there is a line of sight to a satellite, communication can be performed even in an environment where normal communication lines, optical cables, mobile phones, and Wi-Fi cannot be used. There is a high demand for attaching satellite communication devices to vehicles, ships, etc. and communicating while moving. When the communication destination satellite is a geostationary satellite, since the angle between adjacent satellites of the geostationary satellite is as small as a few degrees, it is necessary to direct the beam with high precision so that radio waves do not leak into other satellites. In some cases, the allowable value of the beam angle deviation is a very small value within 0.2 degrees. In that case, highly accurate arrival angle estimation is required. Also, the required beam spread angle is small and the antenna aperture is wide.

[0003] Well-known arrival angle estimation estimates the arrival angle from the amplitude and phase between a plurality of elements of an array antenna. However, since the aperture of the array antenna for satellites is wide, the number of elements is extremely large, and arrival angle estimation using the amplitude and phase between elements is difficult in terms of circuit scale and calculation amount. In many cases, arrival angle estimation in satellite communication is performed by detecting the reception power fluctuation when the reception beam is swung around the satellite.

[0004] Typical methods include conical scan and sequential lobing (Patent Document 1, Non-Patent Document 1). Since the disclosure of this specification is mainly an invention related to sequential lobing, the description of conical scan is omitted, and sequential lobing will be described using FIG. 10.

[0005] The lower part of Figure 10 is a graph illustrating the angle and gain shape of the antenna beam. The horizontal axis represents the angle of deviation from the angle at which the satellite is expected to be located based on past estimation results, and the vertical axis represents the antenna beam gain. The gain shapes of beam 1, where the beam peak is on the negative side of the deviation angle, and beam 2, where the beam peak is on the positive side of the deviation angle, are plotted. In sequential roving, the beam is swung over time to both sides of a 0-degree deviation angle in this manner, and the received power of radio waves from the satellite at each point of the swing is detected.

[0006] Assume that the satellite's angle changes from the expected angle and is now at the angle indicated by "Current Satellite Angle" instead of 0 degrees on the horizontal axis. In that case, a difference in received power, or received power difference [dB], occurs between beam 1 and beam 2 as illustrated.

[0007] The upper part of Figure 10 is a graph with the horizontal axis representing the angle of deviation and the vertical axis representing the gain difference [dB] between beam 1 and beam 2. The gain difference [dB] of the power received by beam 1 and beam 2 is detected, and the corresponding angle of deviation on the horizontal axis is determined from the value on the vertical axis corresponding to that value in the gain difference curve. This is the estimated angle of deviation of the current satellite angle from the angle initially predicted at the start of the estimation (horizontal axis 0).

[0008] Conventional technologies use parabolic antennas, as shown in Patent Document 1, and work well when their direction is mechanically adjusted. However, parabolic antennas have a large volume and are unsuitable for applications such as installation on the roof of a vehicle, making the use of phased array antennas mounted on a flat substrate more practical. Phased array antennas control the direction of the beam by controlling the phase and amplitude between numerous antenna elements on the substrate.

[0009] The beam spread formed by an antenna depends on the antenna aperture; the smaller the aperture, the wider the beam. When a phased array antenna forms a beam at a low elevation angle, the antenna spread as seen from the beam propagation direction is small, so the beam width in the elevation direction is wider at low elevation angles than at high elevation angles. On the other hand, since the total power that the antenna can output is constant, the peak gain decreases as the beam width widens.

[0010] Figure 11 shows an example of the elevation angle dependence of beam gain and shape in a phased array antenna. The horizontal axis represents the elevation angle (shown as a negative value), and the vertical axis represents the beam gain. Note that a coordinate system close to the NED (North-East-Down) system, where the down (Down) is the positive Z-axis, is adopted. Therefore, in this specification, the elevation angle is basically shown as a negative value, but a larger absolute value indicates a higher elevation angle.

[0011] Figure 11 shows the gain shape in the elevation direction of beams at 10-degree intervals from peak elevation angles of -20 degrees to -80 degrees. Both peak gain and beam width change significantly depending on the elevation angle. In particular, the beam width widens extremely at elevation angles below -30 degrees.

[0012] In sequential roving, the beam is generally swung symmetrically to both sides from its original angle, which is 0 degrees of deviation in Figure 10. However, when swinging the beam in the direction of elevation at low elevation angles, the elevation angle changes as a result of the swing, altering the beam gain and beam shape / width. Figure 12(a) shows an example of swinging the beam by 2.3 degrees in both positive and negative directions from -20 degrees. The peak gains of beam 2, which has a peak at -17.7 degrees, and beam 1, which has a peak at -22.3 degrees, are clearly different. Furthermore, as is evident from Figure 11, the beam shape is significantly asymmetrical at low elevation angles, with a stronger tail on the low-elevation side.

[0013] In this case, the gain at the original angle of -20 degrees changes depending on the direction of the swing. In Figure 12(a), the gain is approximately 0.9 dB greater when the beam is swung at -22.3 degrees. In such cases, estimation is possible by using the gain difference curve obtained by subtracting the gain of beam 2 from the gain of beam 1. The thick line in Figure 12(b) (the curve for "gain change: none") is the gain difference curve obtained in this way, with the horizontal axis at 0 and the vertical axis at approximately 0.9 dB.

[0014] On the other hand, the signals received by each element of a phased array antenna are amplified by a low-noise amplifier. The signals from multiple elements are repeatedly combined and amplified into a single signal, which is then down-converted to an intermediate frequency or baseband frequency for power detection. During this process, the signal is amplified multiple times to compensate for circuit losses. Generally, the gain of an amplifier changes with operating temperature. While some compensation is possible by installing temperature sensors, it is difficult to completely suppress gain fluctuations, and some uncompensated, unknown gain fluctuations remain.

[0015] In satellite communications, the received CNR (Carrier-to-Noise Ratio) is often low, around a few dB, meaning the amount of noise in the received power is not negligible. When noise is present in the received power, the change in received power due to beam deflection is reduced by the noise. The amount of reduction in the change depends on the original CNR. The difference in received power [dB] between the two beams in sequential roving also decreases depending on the CNR, but since the original CNR is unknown, it is impossible to determine how much the difference in received power [dB] decreased due to the included noise, and therefore the accurate deflection angle cannot be estimated.

[0016] Therefore, assuming that the noise is basically generated by thermal noise, the thermal noise, which has been measured in advance, is excluded from the received power to obtain only the signal component power for estimation. However, if there is an unknown gain fluctuation, the amount of noise that should be subtracted from the received power will be incorrect.

[0017] Figure 12(b) shows the gain difference curve plotted as a thin line when the CNR is 3.4 dB with the beam pointed at -20 degrees, and there is an unknown gain fluctuation in the range of ±2 dB, after subtracting the incorrect amount of noise. As can be seen from the figure, the slope of the gain difference curve changes, and at the same time, the intercept at a shift angle of 0 degrees changes.

[0018] In satellite arrival angle tracking when a ground station is moving, if the ground station's angle change is gradual and stable tracking is possible, the predicted satellite arrival angle will be approximately 0 degrees of deviation.

[0019] As shown in Figure 12(b), if the value on the vertical axis at horizontal axis 0 changes from the value assumed beforehand, and the cause is an unknown quantity of gain fluctuation, it is impossible to know how it changed. In that case, estimation can only be performed using the thick curve.

[0020] Let's assume the correct value of the satellite misalignment angle is 0 degrees, and the gain difference curve at that point is the top curve, "Gain change: -2dB". Since the received power difference [dB] is measured at approximately 1.3dB on the vertical axis, interpreting this using the thick gain difference curve would result in a misalignment angle of -0.4 degrees, which is significantly different from the true 0 degrees.

[0021] In the next iteration of sequential roving, this incorrect angle will be treated as a 0-degree deviation. However, if the satellite angle has hardly changed, the satellite will actually be around 0.4 degrees, and there will be no satellite at a 0-degree deviation. If this happens repeatedly, even though there is no significant change in the angular relationship between the satellite and the ground station, the detected angle of sequential roving will fluctuate greatly. Furthermore, its average value will not be the correct satellite angle, but an offset value, and the estimated angle of arrival will have a DC bias error. Unlike errors due to random noise, bias errors are difficult to eliminate if their cause (gain fluctuation) is unknown.

[0022] The bias error is the average value of the fluctuating sequential roving estimated angle due to estimation using an incorrect curve. Even in the example in Figure 12(b), the value is less than 0.4 degrees, but as mentioned above, it cannot be ignored when the allowable angle deviation of the beam to the satellite is 0.2 degrees. Angle deviations of the beam to the satellite are not limited to this; they also occur due to random noise and antenna control. It is not possible to allocate almost all of the 0.2 degrees to bias error due to unknown gain fluctuations. [Prior art documents] [Patent Documents]

[0023] [Patent Document 1] Japanese Patent Publication No. 2017-198662 [Non-patent literature]

[0024] [Non-Patent Document 1] Kuramoto, et al., ”Mechanically steered tracking antenna for land mobile satellite communications,” IEEE Antennas and Propagation Society, AP-S International Symposium, vol.3, pp.1314-1317, 1988 [Summary of the Invention] [Problems to be Solved by the Invention]

[0025] Embodiments of the present invention solve such problems and provide an arrival angle estimation device and an arrival angle estimation method capable of estimating the arrival angle without generating or suppressing a bias error even when there are unknown gain fluctuations. [Means for Solving the Problems]

[0026] The arrival angle estimation device of the present embodiment includes an antenna in which at least one of the beam gain and the beam shape changes depending on the angle of the beam, a power detection unit that detects the power of radio waves from a radio wave transmission source received by the antenna based on the beam, a beam control unit that controls the angle of the beam, and a beam is swung in two or more directions from the center at different widths so that the beam gains at the center are equal around the predicted arrival angle of the radio waves, and the arrival angle of the radio waves is estimated based on the power detected by the power detection unit at the beam angles after each swing. [Brief Description of the Drawings]

[0027] [Figure 1] It is a diagram showing a typical embodiment of the present invention. [Figure 2] It is a diagram for explaining the operation and effect of the embodiment of the present invention. [Figure 3] This is a diagram illustrating the different directions of swing in sequential roving. [Figure 4] This figure shows an example of a database for fluctuation range. [Figure 5] This is a diagram illustrating how to create a power ratio-angle shift database. [Figure 6] A diagram showing an example of a power ratio versus angular displacement DB. [Figure 7] This figure shows another embodiment of the present invention. [Figure 8] This is a flowchart of the process according to an embodiment of the present invention. [Figure 9] This diagram illustrates the external connectivity of the angle of arrival estimation device. [Figure 10] This is a diagram to explain sequential roving. [Figure 11] This figure illustrates the elevation angle dependence of the gain shape of a phased array antenna. [Figure 12] This diagram illustrates the problems with conventional methods. [Modes for carrying out the invention]

[0028] Embodiments of the present invention will be described with reference to the drawings. In the following embodiments, only parts that are essentially necessary for the embodiments of the present invention are shown, and parts that are not related to the operation of the embodiments of the present invention are not shown or described.

[0029] Figure 1 shows one of the representative embodiments of the present invention.

[0030] The angle of arrival estimation device 100 receives radio waves from a geostationary satellite 200, which is a radio wave source, using a beam (receiving beam) formed by a phased array antenna 1, and estimates the angle of arrival or direction of arrival of the radio waves by sequential roving. The phased array antenna 1 is an example of an antenna in which at least one of the beam gain and beam shape changes with the beam angle. In this embodiment, a geostationary satellite 200 is assumed as the satellite, but other types of satellites may also be used.

[0031] The angle of arrival estimation device 100 has an angle of arrival estimation unit 2 that estimates the angle of arrival of radio waves. The angle of arrival estimation unit 2 can be composed of circuits such as a CPU (Central Processing Unit), an ASIC (Application Specific Integrated Circuit), and an FPGA (Field Programmable Gate Array), as well as a storage device such as memory. The angle of arrival estimation unit 2 calculates the next angle to which the beam should be pointed and instructs the beam control unit 5. If the polarization of the radio waves output by the geostationary satellite 200 is not circularly polarized, the polarization is also calculated and instructed at the same time.

[0032] The beam control unit 5 receives the instructed beam angle and controls the phase and amplitude of each antenna element of the phased array antenna 1 to form an appropriate receiving beam. In the example in Figure 1, beam 301 is formed first. The direction 301-1 of the gain peak of beam 301 is an angle that is shifted by δd from the expected angle 200-1 (expected angle of arrival of radio waves) of the geostationary satellite 200. In sequential roving, the direction is known with a certain degree of accuracy from the estimation results of the previous roving, and the beam is swung around that angle 200-1.

[0033] When a signal from geostationary satellite 200 is received by the beam 301 formed by the phased array antenna 1, the received signal is subjected to appropriate amplification, filtering, frequency conversion, etc. by the RF / IF unit 3 and converted to an intermediate frequency (IF) or baseband. The output of the RF / IF unit 3 is branched and sent separately to a demodulator (modem) (not shown) for demodulation (see Figure 9 below). The outputs of each element of the phased array antenna may be combined in a separate system for demodulation to avoid the effects of gain fluctuations and gain reductions caused by sequential roving of the beam.

[0034] The output of the RF / IF unit 3 is input to the power detection unit 4. The power detection unit 4 detects the power of the input signal and outputs it to the angle of arrival estimation unit 2.

[0035] As mentioned above, sequential roving requires the beam to be swung to both sides around the predicted angle 200-1 and the received power ratio calculated. In the arrival angle estimation unit 2, the processing selection unit 10 determines whether the beam has been swung to both sides, and if not, the beam angle calculation unit 7 calculates the beam direction when swung to the opposite side.

[0036] The beam angle calculation unit 7 determines the amplitude by referring to the amplitude DB 8 (first database) corresponding to the current predicted angle 200-1 of the geostationary satellite 200. The amplitude δd is determined similarly when the beam is swung to beam 301. Next, the beam is swung to the opposite side to form beam 302. In this embodiment of the present invention, if the beam shape and gain of beam 301 and beam 302 are different and the beam gain of the predicted angle 200-1 of the geostationary satellite 200 is different when swung by the same amplitude, the amplitudes of beam 301 and 302 are changed. The amplitude of beam 302 is δu (≠δd). The direction 302-1 of the gain peak of beam 302 is an angle shifted by δu from the predicted angle 200-1 of the geostationary satellite 200 (the predicted angle of arrival of the radio waves). The gain of beam 301, formed by swinging by δd, in the direction of the predicted angle 200-1 of the geostationary satellite 200 is almost the same as the gain of beam 302, formed by swinging by δu in the opposite direction, in the direction of the predicted angle 200-1 of the geostationary satellite 200. The amplitudes δd and δu are predetermined so that the gain in the direction of the predicted angle is almost the same, and are stored in the amplitude DB 8. The amplitude DB 8 will be described later.

[0037] The power detected by receiving radio waves from the geostationary satellite 200 in the same way with beam 302 is input to the angle of arrival estimation unit 2. The processing selection unit 10 again determines whether it has swung to both sides, and this time, since it has finished swinging to both sides, the angle deviation calculation unit 6 calculates the angle deviation from the predicted angle of the geostationary satellite 200.

[0038] The angular shift calculation unit 6 first calculates the ratio (dB difference) of the received power at beam 301 and the received power at beam 302. If the CNR of the received power is low (for example, if the CNR is less than a predetermined threshold) and the amount of noise included in the received power is not negligible, the unit first removes the noise-equivalent power from the received power, leaving only the signal-equivalent power. Then, it calculates the ratio (dB difference) of the signal-equivalent powers at the two beams. If the power is expressed in dB, the ratio calculation is a subtraction.

[0039] After calculating the signal equivalent power ratio (dB difference), the angular shift calculation unit 6 refers to the power ratio vs. angular shift DB 9 (second database). Details of the power ratio vs. angular shift DB 9 will be described later. From the power ratio vs. angular shift DB 9, parameters showing the relationship between the power ratio and angular shift at the predicted angle azimuth and elevation of the current geostationary satellite 200 are extracted. From the extracted parameters, the angular shift corresponding to the calculated power ratio is calculated to obtain the angular shift relative to the power ratio.

[0040] As described later, sequential roving involves swinging in two directions (elevation and left / right), and the resulting angular displacement is the displacement in one of those directions. This is added to the predicted angle 200-1 of the geostationary satellite 200 to calculate the latest predicted angle of the geostationary satellite 200, which is output as the angle of arrival. Furthermore, the latest predicted angle of the geostationary satellite 200 is sent to the beam angle calculation unit 7 for the next sequential roving.

[0041] Once the angular displacement calculation is complete, the process returns to the processing selection unit 10 and proceeds to the next sequential roving. Sequential roving can only detect angular displacement in the direction the beam was swung. For satellite arrival angle estimation, which requires estimation of azimuth and elevation, one direction is insufficient; two directions must be swung. The two directions may be performed alternately, or the frequency of the direction that is prone to displacement may be increased. The processing selection unit 10 appropriately selects the type of direction to swing for the next sequential roving, and within that type, whether it is up or down for the elevation direction, or left or right for the left-right direction (described later), and instructs the beam angle calculation unit 7. In this way, the arrival angle for each type of direction is estimated.

[0042] This method prevents bias errors caused by unknown gain fluctuations. A brief explanation of why this effect is achieved is provided below.

[0043] Assume that geostationary satellite 200 is at an elevation angle of -20 degrees, and that the CNR is small at 3.4 dB when the beam is pointed towards it. Because the CNR is small, it is necessary to subtract the noise equivalent power from the received power to obtain the signal equivalent power. Figure 2(a) shows the two beam gains when δd = 1.85 degrees and δu = 2.9 degrees so that the beam gains at -20 degrees are approximately the same. If there is an unknown gain fluctuation of -2 dB to 2 dB in the RF / IF section 3, and the gain fluctuation cannot be reflected in the noise equivalent power to be subtracted, the signal equivalent power ratio (dB difference) of the two beams will be as shown by the thin line in Figure 2(b). Similar to Figure 12(b), the slope changes significantly due to the unknown gain fluctuation, but the value on the vertical axis at horizontal axis 0 hardly changes. Therefore, if tracking is performed well and geostationary satellite 200 is at an angle close to the predicted angle of the satellite, the error in the estimated angle in sequential roving is small even if there is an unknown gain fluctuation. Furthermore, even if the satellite is located at a position other than zero angle of deviation, if the rate of change of the true angle of geostationary satellite 200 is small, the angle deviation will converge to zero as the estimation is repeated. If the slope of the gain difference curve is incorrect, there will be some oscillations during the convergence process, but it will converge to zero angle deviation, and no bias error will occur.

[0044] Two types of directions in which the beam is swung in sequential roving according to an embodiment of the present invention will be described. Since the angle of arrival requires the estimation of both azimuth and elevation, sequential roving is usually performed in two types of directions, as shown in Figure 3. In the example in Figure 3, the beam is swung in two orthogonal directions: elevation direction 303 and left-right direction 304. With respect to the predicted angle 200-1 of the geostationary satellite 200, the beam is swung in two opposite directions indicated by arrows in the elevation direction 303, and in two opposite directions indicated by arrows in the left-right direction 304.

[0045] In this embodiment, the elevation direction 303 is the elevation angle used in a normal polar coordinate system, but the other direction 304 is not the azimuth used in a polar coordinate system, but rather left and right. In Figure 3, the azimuth direction 305 is shown separately as a thin line, but on the sphere 404 plane formed by a vector of length 1 in three-dimensional space, the azimuth direction 305 is not orthogonal to the elevation direction 303. As is clear from the figure, the circle formed by the azimuth direction 305 becomes smaller as the elevation angle increases. For example, near the zenith, the azimuth circle becomes very small, and no matter how much you swing in the azimuth direction, the actual direction of the beam hardly changes, making it practically impossible to estimate the angle of arrival in the azimuth direction.

[0046] In this embodiment of the present invention, the antenna is swung not in the azimuth direction, but in the left-right direction 304 which is perpendicular to the elevation direction 303. The left-right direction 304 is a direction on the circle 403 that includes both the vector 401 corresponding to the predicted angle of the geostationary satellite 200 on the sphere 404 in the figure, and the vector 402 which is perpendicular to it and is on the antenna surface 405.

[0047] In sequential roving in the elevation direction 303, the elevation angle is estimated, but when swung in the left-right direction 304, as can be seen from Figure 3, if the deviation angle is not zero, not only the azimuth but also the elevation angle is updated.

[0048] By doing this, it becomes possible to estimate the azimuth and elevation angle of geostationary satellite 200, regardless of its angle.

[0049] Next, we will explain the range DB 8. Figure 4 shows an example of amplitude DB 8. Here, an example of DB in the elevation direction 303 is shown. This is an example of amplitude DB in 1-degree increments for both the azimuth and elevation of the beam angle. Examples of amplitude when swinging up and down in the elevation direction 303 are shown. If the antenna (phased array antenna 1) is circular (rotationally symmetric), the azimuth dimension (the third dimension in the figure) may not be necessary. An example of a rotationally symmetric antenna is one in which the antenna elements are evenly distributed within a circle, but this is not the only example. The amplitude in the left-right direction 304 is often unnecessary for DB purposes, for reasons that will be explained later, so it is not shown. Below, amplitude DB 8 will be explained in detail.

[0050] The amplitude DB 8 can be created by first measuring the antenna gain shape for each azimuth and elevation angle in the elevation and left-right directions, and then pre-determining and storing the amplitude for each direction of oscillation, as well as for each direction of oscillation within each direction, based on these measurements.

[0051] Due to database size limitations, if phased array antenna 1 is nearly rotationally symmetric, the database will be calculated using the elevation angle of the predicted satellite angle; otherwise, the database will be calculated using both azimuth and elevation. While finer intervals for azimuth and elevation of the predicted satellite angle are preferable, database size limitations mean that intervals are generally limited to around 1 degree. Coarser intervals are also acceptable. If the predicted satellite angle falls in between these intervals, the database will be rounded to the nearest azimuth and elevation.

[0052] If the phased array antenna 1 is rotationally symmetric, the change in beam gain when swung left or right will be almost independent of the azimuth, and therefore the swing amplitude DB 8 in the left or right direction will depend only on the elevation angle. Also, for the same reason, if it is rotationally symmetric, the changes in beam shape and gain when swung left and right will be line-symmetric. Even if swung by the same amount in the left or right direction, the beam gain at the predicted angle will be almost equal, so the same value can be used for both swings. Of course, if it is not rotationally symmetric, it is best to swing by different values ​​so that the gain at the predicted angle is almost the same. In that case, DB will also depend on the azimuth.

[0053] Furthermore, in the phased array antenna 1, the peak antenna gain decreases as the elevation angle decreases. However, as mentioned above, sequential roving is estimated from the power ratio (dB difference) when swung to both sides. Therefore, if the amount of gain reduction at the predicted angle when swung to the left or right does not depend on the elevation angle, and only the peak gain decreases, then the amplitude of the swing in the left-right direction can be the same and completely independent of the azimuth, elevation angle, and direction of swing. Of course, if the beam shape in the left-right direction changes in response to the elevation angle, not just the gain, then the amplitude of the swing in the left-right direction can be changed in response to the elevation angle.

[0054] The amplitude of the elevation swing depends only on the elevation angle if the phased array antenna 1 is rotationally symmetric. Different DBs with different amplitudes are created so that the gain at the predicted angle is equal when swinging to a high elevation and when swinging to a low elevation. If the phased array antenna 1 cannot be said to be rotationally symmetric, the amplitude may also depend on the azimuth.

[0055] Even if the phased array antenna 1 is designed with rotational symmetry, the beam shape will vary slightly depending on the direction due to the accuracy of beam control, the antenna's condition, such as substrate temperature, radome shape, and peripheral circuits. Not only is there an imperfection in beam formation, but there are also limitations on the DB size, and since the DB is discrete for each direction and elevation angle, if the predicted angle is in between, the beam shape will be slightly different, and the gain difference at the predicted angle will not be exactly zero. Even if the predicted angle is exactly the value stored in the DB, if the amplitude DB 8 is created depending only on the elevation angle, the gains of the two beams at the predicted angles will not be exactly the same depending on the direction.

[0056] The amplitude DB 8 is designed so that the gain difference [dB] at the predicted angle is as close to zero as possible, but in reality it cannot be made completely zero. In the embodiment of the present invention, even if the gain difference curve changes due to the expected unknown gain fluctuations, with respect to the satellite angular deviation tolerance, for example 0.2 degrees in the above example, the maximum absolute value of the angular deviation at a gain difference of zero is made sufficiently smaller than the tolerance, for example, about 1 / 3 or less.

[0057] Next, we will explain the power ratio versus angular misalignment DB 9.

[0058] Figure 2 shows an example where the beam is swung in two directions of elevation when the elevation angle is -20 degrees, while Figure 2(b) shows the horizontal axis as the displacement angle and the vertical axis as the power ratio (gain difference in dB). When actually estimating the displacement angle from the power ratio obtained by sequential roving, the horizontal axis value is determined from the vertical axis in Figure 2(b). Therefore, it is preferable that the parameters stored in the DB be parameters related to a line inverted vertically and horizontally. The Power Ratio vs. Angle Displacement DB 9 holds parameters related to this line for each azimuth, elevation angle, and type of direction in which sequential roving is performed.

[0059] However, beam geometry cannot be described by a simple equation, and the power ratio versus angular displacement curve that accurately reflects the beam geometry cannot be expressed by a simple equation either. While it would be possible to store the values ​​on the vertical and horizontal axes of a finely discretized curve in a database to reproduce this curve with high accuracy, this would result in an excessively large database size.

[0060] On the other hand, the range of angular displacement that can be estimated by sequential roving is not very wide when the CNR is small, due to the problem that the gain in the direction of the satellite decreases when the beam is swung. Since the gain curve is quite close to a straight line when limited to the vicinity of the central angle, in the embodiment of the present invention, the gain difference curve is approximated by a straight line, and only its intercept and slope are stored in the DB. In this case, the horizontal axis is the gain difference [dB], and the vertical axis is the angular displacement.

[0061] Figure 5(a) shows an example of linear approximation. The solid line is the gain difference curve before approximation, and the dotted line with a small pitch is a straight line approximated by the least squares method near the 0-degree deviation angle of this curve. Figure 5(b) shows an enlarged view of the vicinity of the origin. If linear approximation is simply performed using the least squares method, the approximation line does not pass through the origin because the original curve is not a perfect straight line. For the purposes of the embodiments of the present invention, it is very important that this line passes through the origin. Therefore, the approximation line is shifted so that it passes through the origin as shown by the dashed line, or so that the intercept passing through the horizontal axis 0 is the same as that of the original gain difference curve. The resulting approximation line is a dashed line with a large pitch.

[0062] Figure 6 shows an example of a power ratio versus angular displacement DB 9. An example of DB with the beam swung in the elevation direction (a) and an example of DB with the beam swung in the left-right direction (b) are shown. Both azimuth and elevation are given in 1-degree increments. The intercept and slope are indicated. The intercept and slope determine a straight line where the input is the power ratio [dB] and the output is the displacement angle [deg]. If the antenna is circular (rotationally symmetric), the azimuth dimension (the third dimension in the figure) may be unnecessary. The intercept should be approximately 0, but examples are shown where it is not perfectly 0 due to imperfections in the beam shape or the quantization unit of the amplitude.

[0063] In this way, the Power Ratio vs. Angular Offset DB 9 stores the correspondence between the power ratio and the angle of deviation of the geostationary satellite 200 from the center (predicted angle of geostationary satellite 200) for each elevation angle of the beam angle. In particular, this correspondence is represented by the slope and intercept of a straight line that approximates the curve showing the relationship between the power ratio and the angle of deviation of the geostationary satellite 200 from the center. By creating the Power Ratio vs. Angular Offset DB 9 in this way, the DB size can be reduced without compromising the bias error suppression performance.

[0064] Another embodiment of the present invention is shown in Figure 7. Figure 7 shows only a portion of the arrival angle estimation device 101. The inside of the arrival angle estimation unit 2 is omitted.

[0065] Unlike Figure 1, the power detection unit 11 has a signal power detection unit 12 and a noise power detection unit 13, and performs signal power detection and noise power detection almost simultaneously. Under low CNR conditions, it is necessary to remove the noise power from the detected signal power, but the noise power changes moment by moment. It is difficult to measure the noise power itself that is included in the signal power, but the noise power at that time can be predicted from the received power in a different band or time period where no signal is present.

[0066] The power detection unit 11, for example, branches the signal input from the RF / IF unit 3 and inputs it to the signal power detection unit 12 and the noise power detection unit 13. The signal power detection unit 12 detects the power of the input signal at the frequencies where the signal exists, and the noise power detection unit 13 detects the power of frequencies where the signal does not exist, for example, the power of the guard band. The power detected by the signal power detection unit 12 is essentially the power of the signal plus noise. Both of these outputs are input to the angle of arrival estimation unit 2. In the angle of arrival estimation unit 2, when removing noise power during angle shift calculation, the amount of noise power input from the noise power detection unit 13 is subtracted from the signal power detected by the signal power detection unit 12 to obtain the signal equivalent power.

[0067] Furthermore, in systems where the received signal is packetized and guard intervals are inserted between packets, signal power detection and noise power detection can be performed on the same system, detecting signal power when the signal is being transmitted and noise power when it is not.

[0068] Since the noise is usually random noise, its level fluctuates randomly in accordance with the reception period. Sequential roving is applied to the signal power, and if the angle of arrival estimation device 101 is mounted on a moving object, the angle of the geostationary satellite 200 changes in accordance with the movement. Therefore, measurements cannot be taken for very long periods. On the other hand, since thermal noise is assumed, the average power does not change rapidly. Therefore, it is preferable to output the noise power as an averaged value over a sufficiently longer averaging period than the signal power.

[0069] Because sequential roving swings the beam over time, when swinging to the other direction after detecting power by swinging in one direction, the predicted angle of the geostationary satellite 200 may change due to the movement of the mobile body on which the phased array antenna 1 is mounted. If the amplitude of the swing in each direction changes in increments of 1 degree in azimuth and elevation, the appropriate amplitude at the initial predicted angle may differ from the appropriate amplitude after the change in predicted angle.

[0070] If the amplitude DB 8 is in 1-degree increments, a 1-degree difference will not cause an extreme change in amplitude. Also, since the power ratio versus angle deviation DB 9 may be created for that azimuth / elevation angle amplitude combination, it is best to use the amplitude set for the initial predicted angle that has been optimized as a combination.

[0071] However, if the predicted angle changes, the beam shape will change slightly, and the gain at the predicted angle after the swing will change. If the angle dependence of the beam shape is large, the effect will be large. In such cases, of the two swing directions, it is better to swing in the direction with the greater angle dependence of the beam shape first. That is, of the two directions in which to swing the beam, swing the beam first in the direction in which the amount of change in beam gain at the center (the expected angle of arrival of the radio waves) is larger when the beam angle fluctuates slightly around the beam angle after the swing. When swinging up and down at low elevation angles, the change in beam shape is greater at lower elevation angles. If you swing in the direction of lower elevation angles last, when the predicted angle changes in the elevation direction, the amount of change in the gain shape at the point of swing will be large. On the other hand, if you swing first, you can point the beam before the satellite changes angle, so the effect of the change in gain shape is small. Therefore, when it is judged that the movement in the elevation direction is large at low elevation angles, swinging downwards first will enable more accurate estimation of the angle of arrival.

[0072] The magnitude of the movement in the elevation direction may be determined from the magnitude of the change in the elevation direction of the most recent arrival angle estimation result, or it may be determined based on data from a gyroscope or inertial measurement unit (not shown).

[0073] Figure 8 is a flowchart of a process according to an embodiment of the present invention, and more specifically, a flowchart of a process for executing an angle of arrival estimation method to estimate the direction of a geostationary satellite 200 that is a radio wave source.

[0074] When the beam is swung in any direction from the predicted angle of the geostationary satellite 200, if the gain at the predicted angle of the geostationary satellite 200 differs in the direction of swing when swung by the same amount on both sides, the swing amplitudes δu and δd (≠δu) at which the gain at the predicted angle of the satellite are approximately equal are obtained by referring to the swing amplitude DB 8 (S601).

[0075] Next, the beam is swung by δu in the direction that should be swung by δu (S602), and the received power from geostationary satellite 200 is measured (S603). This is denoted as Pu [dB].

[0076] Next, the beam is swung by δd in the direction that should be swung by δd (S604), and the received power from geostationary satellite 200 is measured (S605). This is denoted as Pd [dB].

[0077] Let Pu'[dB] be the value obtained by removing noise power from Pu, and Pd' be the value obtained by removing noise power from Pd. Calculate the dB difference between them, ΔP = Pu' - Pd'[dB] (S606).

[0078] Referencing the power ratio versus angular deviation DB 9, obtain the gain curve for geostationary satellite 200 at the predicted angle, and calculate the deviation angle from the predicted angle corresponding to ΔP (S607).

[0079] The predicted angle of the satellite is updated to the latest value based on the calculated deviation angle (S608).

[0080] After this, the process returns to step S601 and similarly calculates the displacement angle and updates the predicted satellite angle by swinging the beam in a different direction than the one previously swung in step S601. For example, the beam may be swung in the elevation direction the first time, and then in the left-right direction the second time.

[0081] Figure 9 illustrates the external connectivity of the angle of arrival estimation device 102. Elements equivalent to those in Figure 1 are denoted by the same reference numerals, and detailed explanations are omitted where appropriate.

[0082] The angle of arrival estimation device 102 is used by being incorporated into the satellite communication device 103. Since the satellite communication device 103 is installed on a mobile body such as a car or ship, the relative direction of the geostationary satellite 200 as seen from the antennas (transmitting phased array antenna 21 and receiving phased array antenna 1) changes with the movement of the mobile body. The angle of arrival estimation unit 2 outputs an estimated angle of arrival value, but this angle gradually deviates from the angle of the geostationary satellite 200 due to the subsequent movement of the mobile body. If the frequency of angle of arrival estimation cannot keep up with the angle change determined by the movement of the mobile body, and the error becomes a non-negligible amount with respect to the beam angle tolerance by the time of the next estimation, beam angle calculation and beam control are required to compensate for the gap.

[0083] Therefore, the satellite communication device 103 has an attitude estimation unit 22 that calculates the antenna's attitude from the measurements of the IMU (Inertial Measurement Unit) 26. The attitude estimation unit 22 outputs the attitude and its time change.

[0084] The beam angle calculation unit 7 calculates the beam angle corresponding to the next direction of deflection instructed by the processing selection unit 10, based on the previous arrival angle output by the angle deviation calculation unit 6. However, time elapses from the time the arrival angle is estimated until the next beam angle calculation is completed, and time also lies between the time the beam angle is instructed to the beam control unit 5 and the completion of antenna control so that the beam points to the specified angle. During this time, the angle of the geostationary satellite 200 may change due to the movement of the moving object.

[0085] Therefore, the beam angle calculation unit 7 uses the amount of change in attitude over time output from the attitude estimation unit 22 to correct the change in satellite angle over time from the time the angle of arrival was estimated to the time the beam angle was calculated. Furthermore, it predicts the amount of change in angle over time until the beam is facing the specified angle, corrects the beam angle, and instructs the beam control unit 5 accordingly.

[0086] In this way, regardless of the movement of the moving object, the beam can be swung around the previously estimated angle of arrival as viewed in a global frame such as the NED system. From the difference in received power between the beams swung in these two directions, the angle shift calculation unit 6 estimates the next satellite's angle of arrival.

[0087] Sequential roving can only estimate the angular deviation from the center. If the beam angle is predicted and the beam is swung as described above, and assuming there was no error in the previous angle of arrival, the attitude change output by the attitude estimation unit 22 was correct, the prediction was perfectly correct, and the received power was perfectly correct, the next estimated angular deviation would be 0. In reality, these parameters always have errors, so these are corrected by estimating the angle of arrival.

[0088] The angle shift calculation unit 6 outputs the arrival angle, which is the center angle at that point in time, calculated by adding the estimated angle shift value to the attitude change amount output by the attitude estimation unit 22.

[0089] The center's angle of arrival used in the very first sequential roving is determined by a search, which is not described here. Since the angle of arrival error obtained by the search is larger than the error in sequential roving, the estimated angle of deviation for the first few times will be somewhat large, but in most cases, it will converge to a small error after two estimations of the elevation angle and left / right angle.

[0090] The arrival angle estimation results obtained by sequential roving contain large random errors, which is within the acceptable error range for the transmission beam angle (0.2 degrees), and may not be usable as a transmission beam angle as is. Therefore, the arrival angle tracking unit 23 suppresses these random errors.

[0091] The arrival angle tracking unit 23 tracks the arrival angle output by the arrival angle estimation device 102 using a Kalman filter. While conventional low-pass filters experience delays corresponding to the noise suppression degree, the arrival angle tracking unit 23 receives an attitude change amount from the attitude estimation unit 22, and the Kalman filter's prediction processing is performed using the attitude change amount, enabling noise suppression without delay.

[0092] The arrival angle, noise-suppressed by the arrival angle tracking unit 23, is sent to the transmit beam calculation unit 24. The transmit beam calculation unit 24 receives this and, using the attitude change amount from the attitude estimation unit 22, calculates the predicted beam angle at the time after antenna control is completed, similar to reception, and simultaneously calculates the polarization, sending the azimuth, elevation angle, and polarization information to the transmit beam control unit 25.

[0093] The transmitting beam control unit 25 controls the transmitting phased array antenna 21 to form a beam with a specified azimuth, elevation angle, and polarization.

[0094] Data to be communicated via the geostationary satellite 200 is input and output to the modem 28 via a user interface (not shown) and a higher-layer processing unit (not shown). The data to be transmitted is modulated by the modem 28 and sent to the phased array antenna 21 via the RF / IF unit 27, and transmitted to the geostationary satellite 200 by the beam formed. The received signal received by the phased array antenna 1 is sent to the modem 28 via the RF / IF unit 3 for demodulation and sent to the user interface via the higher-layer processing unit.

[0095] If the polarization required by the specifications is not circular polarization, it is necessary to transmit with the correct polarization. Since it is difficult for the satellite communication device 103 to accurately measure polarization while moving, it is calculated from the attitude, angle of arrival, and also from the azimuth and elevation angles of the geostationary satellite 200 in the NED system, as well as the required polarization specifications. Polarization calculations are performed by the beam angle calculation unit 7 for reception and by the transmission beam calculation unit 24 for transmission.

[0096] The above embodiments were described using sequential roving, where the beam is swung in two directions from the predicted angle of the geostationary satellite 200. However, even with conical scanning, if the antenna gain shape changes depending on the beam direction, the present invention can be applied to rotate the beam around the predicted angle of the geostationary satellite 200 so that the gain at the predicted angle of the geostationary satellite 200 is equal.

[0097] It should be noted that the present invention is not limited to the embodiments described above, and the components can be modified and implemented in practice without departing from the spirit of the invention. Furthermore, various inventions can be formed by appropriately combining the multiple components disclosed in the above embodiments. For example, some components may be deleted from all the components shown in the embodiments. Moreover, components from different embodiments may be appropriately combined.

[0098] [Item 1] An antenna in which at least one of the beam gain and beam shape changes with the beam angle, A power detection unit that detects the power of radio waves from a radio wave source received by the antenna based on the beam, A beam control unit that controls the angle of the beam, An arrival angle estimation unit estimates the arrival angle of the radio wave based on the power detected by the power detection unit at the beam angle at each of the expected arrival angles of the radio wave, by swinging the beam in two or more directions from the center with different amplitudes so that the beam gain at the center is equal, An angle of arrival estimation device equipped with the following features. [Item 2] The aforementioned antenna is a phased array antenna. An angle of arrival estimation device as described in item 1. [Item 3] The number of directions in which the beam is swung is 2. The arrival angle estimation unit swings the beam in opposite directions from the center and estimates the arrival angle of the radio wave from the ratio of the power at the beam angles at each point of the swing. An angle of arrival estimation device as described in item 1 or 2. [Item 4] The direction in which the beam is swung includes at least the direction of elevation, The arrival angle estimation device according to item 3, wherein the arrival angle estimation unit swings the beam in two directions from the center: in the direction in which the elevation angle increases and in the direction in which it decreases. [Item 5] The power detection unit measures the amount of noise expected to be included in the power, The arrival angle estimation unit calculates the ratio of the powers after subtracting the noise amount from the power. An angle of arrival estimation device as described in item 3 or 4. [Item 6] The arrival angle estimation unit includes a database that stores the amplitude of the beam in each of the two directions for each beam angle in the direction of the elevation angle. An angle of arrival estimation device as described in item 4. [Item 7] The arrival angle estimation unit includes a second database that stores the correspondence between the ratio of the power and the angle of deviation of the direction of the radio wave source from the center, for each beam angle in the direction of the elevation angle. An angle of arrival estimation device as described in item 6. [Item 8] The aforementioned correspondence is the slope and intercept of a straight line that approximates the curve showing the relationship between the ratio of the powers and the angle of deviation of the direction of the radio wave source from the center. Angle of arrival estimation device as described in item 7. [Item 9] The value of the ratio when the deviation angle of the straight line approximated with respect to the curve is 0 degrees is equal to the value of the ratio when the deviation angle of the curve is 0 degrees. Angle of arrival estimation device as described in item 8. [Item 10] The aforementioned radio wave source is a satellite. The aforementioned antenna is mounted on a mobile satellite earth station for communication via the aforementioned satellite. An angle of arrival estimation device as described in any one of items 1 to 9. [Item 11] Of the two directions in which the beam is swung, the beam is first swung in the direction in which the amount of change in the central beam gain is larger when the beam angle fluctuates slightly around the beam angle after the beam has been swung. An angle of arrival estimation device as described in item 4. [Item 12] Controlling the beam angle of an antenna in which at least one of the beam gain and beam shape changes with the beam angle, Based on the beam, the power of the radio waves from the radio wave source received by the antenna is detected. The expected arrival angle of the radio waves from the radio wave source is used as the center, and the beam is swung in two or more directions from the center with different amplitudes so that the beam gain is equal at the center. The arrival angle of the radio waves is then estimated based on the power detected at the beam angles at each of the swung points. Arrival angle estimation method. [Explanation of Symbols]

[0099] 1 Phased array antenna 2 Arrival angle estimator 3 IF section 4 Power detection unit 5. Beam Control Unit 6 Calculation section 7. Beam Angle Calculation Unit 10 Processing Selection Unit 11 Power detection unit 12 Signal power detection unit 13 Noise Power Detection Unit 21 Phased Array Antenna 22 Posture estimation section 23. Tracking section of the arrival angle 24 Transmit Beam Calculation Unit 25 Transmit beam control unit 27 IF section 28 Modems 100 Arrival angle estimation device 101 Arrival angle estimation device 102 Arrival angle estimation device 103 Satellite communication equipment 200 geostationary satellites 301 Beam 302 Beam 303 Elevation direction 304 Left / right direction 305 Azimuth direction 401 Vectors 402 Vectors 403 yen 404 sphere 405 Antenna surface

Claims

1. An antenna in which at least one of the beam gain and beam shape changes with the beam angle, A power detection unit that detects the power of radio waves from a radio wave source received by the antenna based on the beam, A beam control unit that controls the angle of the beam, An arrival angle estimation unit estimates the arrival angle of the radio wave based on the power detected by the power detection unit at the beam angles at the predicted arrival angle of the radio wave, by swinging the beam in two or more directions from the center with different amplitudes so that the beam gain is equal at the center, and swinging the beam in each direction. An angle of arrival estimation device equipped with the following features.

2. The aforementioned antenna is a phased array antenna. An angle of arrival estimation device according to claim 1.

3. The number of directions in which the beam is swung is 2. The arrival angle estimation unit swings the beam in opposite directions from the center and estimates the arrival angle of the radio wave from the ratio of the power at the beam angles at each point of the swing. An angle of arrival estimation device according to claim 1.

4. The direction in which the beam is swung includes at least the direction of elevation, The arrival angle estimation device according to claim 3, wherein the arrival angle estimation unit swings the beam in two directions from the center: in the direction in which the elevation angle increases and in the direction in which it decreases.

5. The power detection unit measures the amount of noise expected to be included in the power, The arrival angle estimation unit calculates the ratio of the powers after subtracting the noise amount from the power. The device for estimating the angle of arrival according to claim 3.

6. The arrival angle estimation unit includes a database that stores the respective amplitudes of the swing of the beam in the two directions for each beam angle in the direction of the elevation angle. The device for estimating the angle of arrival according to claim 4.

7. The arrival angle estimation unit includes a second database that stores the correspondence between the ratio of the power and the angle of deviation of the direction of the radio wave source from the center, for each beam angle in the direction of the elevation angle. The arrival angle estimation device according to claim 6.

8. The aforementioned correspondence is the slope and intercept of a straight line that approximates the curve showing the relationship between the ratio of the powers and the angle of deviation of the direction of the radio wave source from the center. The device for estimating the angle of arrival according to claim 7.

9. The value of the ratio when the deviation angle of the straight line approximated with respect to the curve is 0 degrees is equal to the value of the ratio when the deviation angle of the curve is 0 degrees. The arrival angle estimation device according to claim 8.

10. The aforementioned radio wave source is a satellite. The aforementioned antenna is mounted on a mobile satellite earth station for communication via the aforementioned satellite. An angle of arrival estimation device according to claim 1.

11. Of the two directions in which the beam is swung, the beam is first swung in the direction in which the amount of change in the central beam gain is larger when the beam angle fluctuates slightly around the beam angle after the beam has been swung. The device for estimating the angle of arrival according to claim 4.

12. Controlling the beam angle of an antenna in which at least one of the beam gain and beam shape changes with the beam angle, Based on the beam, the power of the radio waves from the radio wave source received by the antenna is detected. The expected arrival angle of the radio waves from the radio wave source is used as the center, and the beam is swung in two or more directions from the center with different amplitudes so that the beam gain is equal at the center. The arrival angle of the radio waves is then estimated based on the power detected at the beam angles at each of the swung points. Arrival angle estimation method.

Images